IndustryIROPS first
Disruption is where the money is
A single irregular operations event concentrates rebooking cost, compensation, crew disruption and reputational damage, which makes recovery the highest-value target.
Program 15 · Domain specialisation
SCIKIQ Certified Data and AI EngineerAirlines
An airline is a scheduling problem wearing a brand. Everything — revenue, cost, crew legality, customer experience — is downstream of whether the aircraft, the crew and the passenger arrive at the same gate at the same time. This specialisation is built around disruption, because that is where the sector’s money and reputation are actually decided.
- 80 hrstaught
- 40hands-on AI labs
- 8assessment tests
- +20 hrs2 E2E capstones
IROPSWhere cost, compensation and reputation concentrate in a single event
Fare rulesA natural-language corpus governing revenue, read by very few people
AncillaryThe margin line that grew fastest and is least automated
Crew legalityA hard constraint no recommendation may ever violate
01
Part 1 · The business
The value chain, the people, the trends
The chain runs from a schedule published months out to a settlement completed months later. Part 1 maps it and names the people who hold the network together when it breaks.
The airline value chain
Six stages from network planning to settlement. Select one.
01
Plan the network
Route and schedule planning, fleet assignment, slots and seasonal capacity.
Decisions made here
- Which routes do we fly?
- What aircraft goes where?
- How much connection buffer do we build in?
Where it breaks today
- Schedule robustness assessed after the season, not during planning
- Route decisions separated from the operational cost they create
- Competitive and slot intelligence assembled manually
What is measured
- Load factor
- ASK and RASK
- Schedule robustness
Where AI landsSchedule robustness simulation, route case assembly and slot document analysis.
Load factorASK and RASKSchedule robustness
02
Price & manage revenue
Fare filing, fare rules, revenue management, overbooking and demand forecasting.
Decisions made here
- What do we charge?
- How many seats do we protect?
- What does this fare rule actually permit?
Where it breaks today
- Fare rules written in a dense text format almost nobody can read
- Rule interpretation done by a handful of experts nearing retirement
- Forecast overrides applied without a record of why
What is measured
- RASK
- Yield
- Forecast accuracy
Where AI landsFare-rule interpretation with citation, demand anomaly explanation and override capture.
RASKYieldForecast accuracy
03
Sell & distribute
Direct and indirect distribution, GDS, NDC offers, ancillaries and partner selling.
Decisions made here
- What offer do we show?
- Which ancillary do we attach?
- Is this channel earning its cost?
Where it breaks today
- Offer construction constrained by legacy distribution semantics
- Ancillary attachment driven by static rules rather than context
- Channel economics that are hard to attribute
What is measured
- Ancillary revenue per passenger
- Direct share
- Conversion
Where AI landsContext-aware ancillary offers, NDC offer assembly and channel economics attribution.
Ancillary revenue per passengerDirect shareConversion
04
Operate the day
Operations control, turnaround, crew tracking, ATC and airport coordination.
Decisions made here
- Will this flight go on time?
- What breaks next?
- Which aircraft do we swap?
Where it breaks today
- Delay causes coded inconsistently after the fact
- Turnaround problems visible only once they are already late
- Decisions made in a control centre under time pressure with partial data
What is measured
- On-time performance
- Turnaround time
- Cancellations
Where AI landsDelay propagation prediction, turnaround risk flagging and swap option generation.
On-time performanceTurnaround timeCancellations
05
Recover from disruption
IROPS: cancellations, rebooking, crew legality, hotels, compensation and communication.
Decisions made here
- Who do we rebook where?
- Is this crew legal?
- What do we owe this passenger?
Where it breaks today
- Rebooking done by queue position rather than by passenger value or need
- Compensation entitlement assessed inconsistently across events
- Passengers informed last, after social media
What is measured
- Recovery time
- Compensation cost
- Misconnect rate
Where AI landsRebooking option generation within crew and legal constraints, entitlement determination.
Recovery timeCompensation costMisconnect rate
06
Serve, settle & retain
Customer service, claims, interline and MRO settlement, loyalty and retention.
Decisions made here
- Is this claim valid?
- What is this passenger worth?
- Did the partner bill us correctly?
Where it breaks today
- Claim handling that re-reads the same regulation every time
- Interline and MRO invoices reconciled by sampling
- Loyalty value invisible at the moment of a service decision
What is measured
- Claim cycle time
- Cost per claim
- Retention
Where AI landsEntitlement determination with regulation cited, invoice reconciliation and value-aware service.
Claim cycle timeCost per claimRetention
Who you are building for
On a disrupted day the operations controller is making decisions in minutes that cost millions. Select one to light the stages they own.
What the sector is seeing
A high-volume, thin-margin industry where disruption is the concentrated cost event.
IndustryRules as language
Fare rules are an untapped language problem
Revenue-governing rules sit in a dense filed text format that very few people can read, which is exactly the shape of problem language models handle well.
IndustryContext-aware offers
Ancillary is the margin line
Ancillary revenue has grown into a primary contribution source while attachment logic in most carriers remains static and rule-based.
IndustryEvidenced entitlement
Compensation regimes are converging on strictness
Passenger-rights regimes increasingly require consistent, evidenced entitlement decisions, which punishes manual and inconsistent claim handling.
IndustryLegality or nothing
Crew legality is the constraint that binds
Any recovery recommendation that ignores flight-time limitations or agreement terms is not a recommendation, and controllers will discard the tool that makes them.
IndustryOffer and order
Distribution is being rebuilt around offers
The shift toward offer-and-order distribution changes what a carrier must be able to construct and price in real time.
02
Part 2 · Data and meaning
Domains, ontology, taxonomy and governed terms
Airline data is unusually good and unusually siloed: the PSS knows the passenger, operations knows the aircraft, crew systems know legality, and the fare corpus governs what any of it may cost. The engineering is joining them fast enough to matter on a disrupted day.
The data landscape
Six families. The fare and regulation corpora are the most valuable and least exploited.
Passenger service system
Bookings, passengers, itineraries, tickets and check-in.
PSS, reservations, DCSSchedule & operations
Schedules, flight legs, movements, delays and delay codes.
Ops control systems, movement feedsCrew
Rosters, qualifications, flight-time limitations and agreement terms.
Crew management systemsFares & rules
Filed fares, fare rules, taxes and ancillary catalogues.
Fare filing, distribution feedsEngineering & MRO
Aircraft configuration, defects, maintenance and part history.
MRO systems, technical logsService, claims & loyalty
Contacts, claims, compensation, interline and loyalty state.
CRM, claims platforms, loyaltyThe semantic layer
An airline agent that cannot tell a leg from a segment, or a passenger from a booking, will produce recovery options that look sensible and cannot be flown. These are the domains, the graph and the definitions it must be grounded on before it goes near a disrupted day.
Data domains
Six subject areas, each with the entities it holds, its critical data elements and the function accountable for it.
Schedule & network
What is planned to fly, with what aircraft, from where.
RouteFlightLegSlotStation
- Critical data elements
- Flight number; Scheduled times; Equipment type
- Owned by
- Network planning
Passenger & booking
Who is travelling, on what itinerary, under what ticket.
PassengerBookingItinerarySegmentTicket
- Critical data elements
- Record locator; Ticket number; Fare basis
- Owned by
- Commercial & PSS
Fare & rule
What was charged and what the conditions permit.
FareFare ruleTaxAncillary
- Critical data elements
- Fare basis code; Rule category; Ancillary code
- Owned by
- Pricing & revenue management
Operations
What actually happened to the aircraft and the flight.
MovementDelayCancellationTurnaround
- Critical data elements
- Actual times; Delay code; Delay minutes
- Owned by
- Operations control
Crew
Who can legally operate the flight, and until when.
Crew memberPairingDutyQualification
- Critical data elements
- Duty start and end; Flight-time accrued; Qualification validity
- Owned by
- Crew management
Service & entitlement
What the passenger experienced and what they are owed.
Disruption eventClaimCompensationLoyalty member
- Critical data elements
- Delay at arrival; Entitlement basis; Compensation amount
- Owned by
- Customer service
Ontology & knowledge graph
The entities and the typed relationships between them — what the agent traverses instead of guessing joins. Select any entity.
Core entityEventReference / classification
Core entity
Passenger
The person travelling.
- Passenger holds Booking
Core entity
Booking
The reservation record holding an itinerary.
- Booking contains Itinerary
- Booking priced by Fare
- Passenger holds Booking
Core entity
Itinerary
The ordered set of segments a passenger holds.
- Itinerary composed of Segment
- Booking contains Itinerary
Core entity
Segment
One marketed portion of the journey.
- Segment operated as Leg
- Itinerary composed of Segment
Core entity
Leg
One operated flight between two stations.
- Leg scheduled as Flight
- Leg flown by Aircraft
- Leg departs from Station
- Segment operated as Leg
- Crew pairing operates Leg
- Delay affects Leg
Reference / classification
Flight
The scheduled service a leg operates under.
- Leg scheduled as Flight
Core entity
Aircraft
The airframe assigned to the leg.
- Leg flown by Aircraft
Reference / classification
Station
An airport the network serves.
- Leg departs from Station
Core entity
Crew pairing
The sequence of duties assigned to a crew.
- Crew pairing operates Leg
- Delay propagates to Crew pairing
Reference / classification
Fare
The priced conditions the ticket was sold under.
- Booking priced by Fare
- Claim assessed against Fare
Event
Delay
A departure or arrival later than scheduled.
- Delay affects Leg
- Delay propagates to Crew pairing
- Claim arises from Delay
Event
Claim
A passenger request for compensation or redress.
- Claim arises from Delay
- Claim assessed against Fare
Taxonomies
The classification hierarchies that make records comparable across systems.
Network hierarchy
- Region
- Station
- Route
- Flight
Europe → LHR → LHR-JFK → XX117
Delay cause
- Group
- Code
- Attributable
Technical → Aircraft defect → Carrier attributable
Fare rule category
- Category
- Provision
- Condition
Changes → Permitted before departure → Fee plus fare difference
Governed business terms
The definitions an agent must use rather than invent. Most wrong answers in this industry are a term used loosely.
- Segment vs leg
- A segment is what was sold; a leg is what is flown. A one-segment sale can be two legs, and recovery logic that confuses them rebooks the wrong people.
- Passenger vs booking
- A booking can hold several passengers. Rebooking at booking level splits or strands parties who were travelling together.
- Misconnect
- A passenger whose inbound arrival leaves insufficient time for the onward segment — the unit disruption is measured in.
- Minimum connect time
- The station-specific minimum a connection needs to be viable, and the threshold a misconnect is judged against.
- Flight-time limitation
- The regulated cap on crew duty and flying hours. A hard constraint: an option that breaches it is invalid, not merely worse.
- Attributable delay
- A delay assigned to a cause the carrier is responsible for, which is what compensation entitlement usually turns on.
- Fare basis
- The code identifying the fare and therefore the rules governing changes, refunds and carriage.
- Involuntary rebooking
- Rebooking caused by the carrier rather than requested by the passenger, carrying different entitlement.
03
Part 2 · The architecture
How it is actually built
The layers a deployment needs, what each one holds, and what the engineer owns there.
The reference architecture
Seven layers, with crew legality and passenger-rights regulation treated as hard constraints.
01/07
PSS, ops & crew sources
Reservations, operations, crew and the fare and regulation corpora, read under time pressure.
Takes in
Nothing. This is the upstream edge.
Hands on
Passenger and itinerary state, live movements, authoritative crew legality and the fare corpus.
SystemWhat it doesKey business entities
Passenger service system
Reservations, tickets, itineraries and check-in. Holds who is travelling on what, and is the system every recovery action ultimately has to write back into.
PassengerBookingItineraryTicket
Operations control systems
Schedules, movements, delays and cancellations in near real time. The picture a controller is acting on, and the source of the delay coding everything downstream depends on.
FlightLegMovementDelay
Crew management
Rosters, qualifications, duty state and the flight-time limitation calculation. The authoritative answer on legality, which is never reimplemented elsewhere.
Crew memberPairingDutyQualification
Fare filing and distribution
Filed fares, fare rules, taxes and ancillary catalogues, in a dense positional text format that governs what may be charged and changed.
FareFare ruleTaxAncillary
Engineering and MRO
Aircraft configuration, defects, maintenance status and part history. Determines whether an airframe is available to a recovery option at all.
AircraftDefectMaintenance event
Service, claims and loyalty
Contacts, disruption claims, compensation decisions and loyalty state. Where entitlement is determined and where passenger value becomes visible to a service decision.
ClaimCompensationLoyalty member
What the engineer owns here
Negotiating a read path fast enough to be useful inside a live recovery decision.
Where this layer fails
Reimplementing the flight-time limitation calculation because the crew system is slow to call. The moment your copy disagrees with the system of record, every option you produce is unusable.
02/07
Ingestion & itinerary assembly
Event ingestion plus assembly of passengers, itineraries, legs and connections.
Takes in
Read-mostly access paths, change feeds and document streams from the systems of record.
Hands on
Replayable, resolved, quality-checked records with their lineage back to the source row.
SystemWhat it doesKey business entities
Change data capture
Reads the source’s own change log rather than polling it, so the platform sees every state a record passed through instead of only where it ended up.
Change eventOperation typeCommit timestamp
Landing and raw zone
Every extract kept as it arrived, immutable and timestamped. The thing you replay from when a downstream definition turns out to have been wrong.
Raw extractBatchWatermark
Document processing
Layout-aware extraction over the unstructured half of the estate: sectioning, tables, signatures, and the page reference every later citation depends on.
DocumentSectionPage reference
Entity resolution
Decides that two records are the same real-world thing, with a survivorship rule and a confidence, so downstream joins are a decision rather than an assumption.
Match candidateSurvivorship ruleCluster
Data quality and contracts
Schema, freshness and volume expectations asserted at the boundary, so a bad load fails loudly here rather than quietly two layers later.
ContractExpectationViolation
What the engineer owns here
Building the connection graph correctly — a misconnect the system cannot see is invisible.
Where this layer fails
Snapshot loads instead of change capture. It looks identical in a demo and it silently loses every intermediate state, which is exactly what an audit later asks you for.
03/07
Network model & semantics
The graph joining passenger, itinerary, leg, aircraft, crew pairing and station.
Takes in
Replayable, resolved records with lineage from ingestion.
Hands on
Typed, classified, defined data an agent can be grounded on without inferring meaning.
SystemWhat it doesKey business entities
Business glossary
One agreed definition per term, owned by a named person, so the number an agent quotes means what the business means by it.
TermDefinitionOwnerSteward
Ontology and knowledge graph
The typed entities and relationships the domain actually has, so an agent can traverse "which customers are exposed to this" rather than guess from adjacent text.
EntityRelationshipNodeEdge
Semantic and metrics layer
Metrics defined once, in one place, with their filters and grain. Removes the class of error where the agent computed something plausible and wrong.
MetricDimensionGrainFilter
Classification and policy
Sensitivity labels and the purpose each classification permits, applied at the field level and inherited by everything downstream.
ClassificationPurposePolicy
Lineage and catalogue
Where a value came from and what it touched on the way. The layer that makes an answer defensible rather than merely correct.
Lineage edgeDatasetColumn
What the engineer owns here
One definition of a misconnect and of on-time, agreed across commercial and operations.
Where this layer fails
Letting the model infer meaning from column names. It will, it will be plausible, and nobody will notice until the number reaches a regulator or a board pack.
04/07
Retrieval & grounding
Retrieval over fare rules, passenger-rights regulation, contracts of carriage and manuals.
Takes in
Typed, classified, defined data from the governance layer.
Hands on
Ranked, permission-trimmed, citable evidence scoped to the caller and the moment.
SystemWhat it doesKey business entities
Hybrid index
Vector and lexical retrieval together, because exact identifiers, codes and clause numbers are the thing semantic search is worst at.
ChunkEmbeddingTermScore
Chunking and metadata
Splits on the document’s own structure and carries effective dates, version and source into every chunk, so a retrieved passage knows when it was true.
ChunkEffective dateVersionSource reference
Security trimming
Applies the caller’s permissions inside the query rather than filtering results afterwards, so the model never sees what the user may not.
PrincipalEntitlementFilter
Graph retrieval
Traverses the knowledge graph for questions that are joins rather than similarity — exposure, genealogy, ownership, causation.
PathHopTraversal
Reranking and citation
Reorders candidates on relevance and returns the passage identifier behind every sentence, so the answer can be checked rather than trusted.
CandidateRankCitation
What the engineer owns here
Making a fare rule answerable in natural language with the rule text cited.
Where this layer fails
Filtering results after retrieval instead of constraining the query. The model has already seen the rows you removed, and it will use them.
05/07
Agent & orchestration
Agents for recovery option generation, entitlement determination and offer construction.
Takes in
Ranked, permission-trimmed, citable evidence from retrieval.
Hands on
Actions taken or proposed, each with the evidence, the identity and the trace behind it.
SystemWhat it doesKey business entities
Agent runtime
Plans, calls tools and holds the loop. Where the step budget, the timeout and the stopping condition are enforced rather than hoped for.
TaskPlanStepTrace
Tool and MCP servers
Typed, permissioned tools with declared schemas. An agent’s real capability surface is this list, which is why the list is a security artefact.
ToolSchemaScopeInvocation
State and memory
Durable task state with checkpoints, so a run that dies mid-way resumes instead of restarting and re-doing side effects.
CheckpointSessionMemory record
Human-in-the-loop gates
Approval steps on the actions that need one, carrying enough context for the approver to actually decide rather than rubber-stamp.
ApprovalThresholdDecision
Workflow integration
Writes back into the systems people already work in, under a service identity with its own audit trail.
Work itemWrite-backActor
Crew legality servicethis domain adds
The authoritative flight-time limitation check, called rather than reimplemented, and used to build the feasible set before any option is generated.
DutyFlight-time accruedLegality verdict
What the engineer owns here
Generating options a controller can act on, never actions the agent takes alone.
Where this layer fails
Generating options first and filtering illegal ones afterwards. Under time pressure a filter gets bypassed; a feasible set built legally cannot be.
06/07
Legality, rights & guardrails
Crew legality enforcement, passenger-rights compliance, safety boundary and audit.
Takes in
Proposed actions and drafted responses from the agent layer.
Hands on
Permitted, grounded, logged output — or a refusal with a stated reason.
SystemWhat it doesKey business entities
Policy enforcement
The rules that decide whether an action is permitted at all, evaluated before the action and independently of the model that proposed it.
PolicyDecisionObligation
Input and output screening
Injection detection on the way in, and on the way out the checks for leakage, unsupported claims and content the domain forbids.
PromptResponseViolation
Groundedness checking
Verifies each assertion resolves to retrieved evidence, and fails the response rather than shipping the sentence that does not.
AssertionSupportVerdict
Model risk management
Model inventory, intended use, validation evidence and the sign-off that lets a model be used for a purpose. Not optional in a regulated estate.
ModelIntended useValidationApproval
Audit log
Immutable record of what was asked, retrieved, decided and done — the artefact a reviewer reads when they do not take your word for it.
Audit entryActorEvidence reference
Passenger-rights compliancethis domain adds
Determines entitlement against the governing regulation with article and version cited, to a standard an auditor rather than a customer would accept.
EntitlementRegulation articleEvidence
What the engineer owns here
Guaranteeing no recommendation can breach flight-time limitations or agreement terms.
Where this layer fails
A plausible wrong entitlement answer. This is a legal determination, so being approximately right is a regulatory exposure rather than a service miss.
07/07
Evaluation & observability
Option quality, entitlement accuracy, latency under disruption and cost per decision.
Takes in
Permitted, grounded, logged output from the guardrail layer.
Hands on
Measured quality, cost and latency — and the evidence to change any of the three.
SystemWhat it doesKey business entities
Evaluation harness
Held-out sets and graded runs on every change, so a prompt edit is a measured change rather than a hopeful one.
Eval setCaseScoreBaseline
Regression gate
Blocks a release when quality drops, in CI, on the same evidence for everyone. The difference between a system and a demo.
ThresholdRunGate decision
Tracing
End-to-end spans across retrieval, model and tool calls, so a bad answer can be opened and read rather than argued about.
TraceSpanLatencyToken count
Cost and capacity
Cost per task and per tenant, against throughput and quota. The number that decides whether the pilot can become the rollout.
Cost per taskQuotaThroughput
Drift and feedback
Watches quality against production traffic rather than the test set, and routes real corrections back into the eval suite.
Drift signalFeedbackIncident
What the engineer owns here
Proving the system still responds inside the decision window on the worst day, not the average one.
Where this layer fails
Evaluating once, before launch. Quality moves with the data, the model and the traffic, and a system with no live measurement has no idea which of the three moved.
Typical tech stack
Models
- Frontier models for rule and regulation reasoning
- Small models for high-volume triage
- Forecast models for demand and delay
Grounding
- Fare rule and regulation index
- Contract of carriage corpus
- Network knowledge graph
Agents
- Recovery option agents
- MCP over PSS and operations
- Constraint solver tools
Data
- Real-time movement ingestion
- Itinerary and connection assembly
- Crew legality data
Assurance
- Option quality evals
- Entitlement accuracy measurement
- Legality regression tests
Runtime
- Low-latency deployment for ops control
- Surge capacity for disruption days
- Degraded-mode behaviour
Constraints that shape the design
Crew legality is absolute
Flight-time limitations and agreement terms are hard constraints. An option that breaches them is not a lesser option; it is an invalid one, and must be unreachable rather than discouraged.
The worst day is the design point
Systems are needed most when disruption spikes load. Capacity and latency must be designed for the storm, not for the median Tuesday.
Passenger rights are a legal determination
Compensation entitlement is regulated and evidenced. A plausible wrong answer is a regulatory exposure, not a customer-service miss.
Nothing touches airworthiness
Maintenance and safety decisions sit behind a regulated boundary. The agent supports planning and administration around it, never the airworthiness decision itself.
04
Part 3 · Getting it done
Where the business and the technology meet
The highest-value use cases cluster around the disrupted day, because that is where decisions are fast, expensive, constrained and currently manual.
Use cases, mapped to the chain and the architecture
Filter by stage or by earned autonomy. Selecting a use case jumps the value chain and the architecture to the stage and layer it depends on.
Value chain stage
Autonomy earned
12 use cases
No use cases match that combination.
The delivery playbook
How an airline deployment actually runs.
01
Week 1 — settle the legality boundary
Agree in writing that crew legality and airworthiness are hard constraints the agent cannot reach past, and confirm it architecturally. Nothing else can be discussed until this is.
02
Week 2 — build the connection graph and test it
Assemble passengers, itineraries, legs and connections and measure how well it reflects reality. A misconnect the system cannot see is a use case that cannot exist.
03
Weeks 3-6 — start with the fare and regulation corpora
Rule and entitlement retrieval delivers value fast, touches no operational system and builds the citation discipline every later use case depends on.
04
Weeks 7-9 — take on recovery with the worst day as the design point
Build option generation against a replayed severe disruption, not a normal day. Latency and capacity under surge are the acceptance criteria.
05
Weeks 10-12 — degraded mode, audit and handover
Prove behaviour when a source system is unavailable mid-recovery, then hand over runbook, legality regression tests and the entitlement audit trail.
The curriculum
8 modules, 80 taught hours, 40 hands-on labs and 8 assessments — every lab provisioned and graded by the SCIKIQ Agentic AI Playground. Open a module to see its labs.
01The SCDAI delivery model in an airlineLanding in an operation where the schedule outranks every other consideration.10 hrs · 5 labs
Delivery mandateCommercial vs operations splitValue sizing on disruptionStakeholder mapController adoptionThin-slice scoping
01
Decision selectionScore four airline decisions on value, cadence and data readiness.
02
Value sizingSize a misconnect, a cancellation and a compensation claim.
03
Graph auditAssess whether itinerary and connection data support a recovery use case.
04
Stakeholder mapMap operations control, crew, revenue management, service and safety.
05
Thin sliceScope a single-station disruption slice with a measurable outcome.
Assessment 01 — Operating model & scoping
02Airline data foundationsPSS, movements, crew, fares and the corpora that govern what any of it may cost.10 hrs · 5 labs
PSS and booking dataMovement and delay codingCrew data and legalityFare and rule formatsMRO dataItinerary and connection assembly
01
PSS extractionExtract passengers, itineraries and tickets into a working model.
02
Delay normalisationNormalise inconsistent delay coding into an analysable set.
03
Connection graphAssemble the connection graph and measure how well it matches reality.
04
Fare corpus ingestionIngest filed fare rules with positional structure preserved.
05
Crew dataModel rosters, qualifications and flight-time limitation inputs.
Assessment 02 — Airline data landscape
03Context engineering for airline languageCodes, times, time zones and regulation — where precision is not negotiable.10 hrs · 5 labs
Station and carrier codesTime zones and UTC disciplineStructured outputsUncertaintyRefusal on legality questionsContext budget
01
Code resolutionResolve ambiguous station, carrier and equipment codes.
02
Time disciplinePrevent time-zone and day-boundary errors in itinerary reasoning.
03
Structured optionsEmit recovery options as a typed, validated structure.
04
Legality refusalMake the agent refuse to assert legality rather than estimate it.
05
Context budgetCut context cost on a high-volume disruption agent.
Assessment 03 — Context engineering
04Grounding on fare rules and passenger rightsTwo corpora that govern revenue and liability, and that almost nobody can read.10 hrs · 5 labs
Fare rule structureRegulation retrievalContract of carriageNetwork graph traversalCitationsRetrieval evals
01
Fare rule retrievalAnswer a fare-rule question with the exact rule text cited.
02
Regulation retrievalRetrieve the governing passenger-rights provision for a scenario.
03
Interline agreementsResolve which agreement governs a disrupted interline itinerary.
04
Citation contractCite rule, article and version behind every determination.
05
Retrieval evalsBuild an expert-labelled rule question set and measure accuracy.
Assessment 04 — Retrieval & grounding
05Disruption agent designOption generation under hard constraints, at speed, on the worst day of the year.10 hrs · 5 labs
Workflow vs agentConstraint handlingSolver integrationOption rankingLatency budgetsEscalation
01
Recovery agentGenerate ranked recovery options with cost and passenger impact.
02
Hard constraintsMake crew-illegal options unreachable rather than merely ranked low.
03
Solver toolsGive the agent a constraint solver and use it for what it is good at.
04
Latency budgetMeet a control-centre latency budget under simulated surge.
05
EscalationDefine when the agent must stop and hand over, and prove it does.
Assessment 05 — Agent design for disruption
06Integrating with PSS, operations and crewRead from systems built in another era, write only where a human approved it.10 hrs · 5 labs
MCP over PSS and opsReal-time movement feedsCrew system integrationApproval gatesIdentityDegraded mode
01
MCP over PSSExpose passenger and itinerary lookups as typed read tools.
02
Movement ingestionConsume a real-time movement feed with ordering guarantees.
03
Crew legality toolCall the authoritative legality check rather than reimplementing it.
04
Approval gateApply a rebooking only after a controller approves it.
05
Degraded modeDefine and test behaviour when a source system fails mid-recovery.
Assessment 06 — Systems integration
07Safety boundary, rights and securityThe module that decides whether safety and legal will let this near the operation.10 hrs · 5 labs
Airworthiness boundaryCrew legality assurancePassenger-rights complianceOWASP LLM Top 10Injection via customer textIncident response
01
Safety boundaryDocument how the agent sits outside airworthiness decisions.
02
Legality regressionBuild a regression suite that proves no illegal option is emitted.
03
Rights complianceEvidence an entitlement determination to an auditor’s standard.
04
Injection defenceDefend against instructions embedded in a passenger claim.
05
Red teamRun an OWASP LLM Top 10 pass over the recovery pipeline.
Assessment 07 — Safety, rights & security
08Production, surge and cost per decisionThe system is needed most exactly when load is highest. Design for that day.10 hrs · 5 labs
Eval suitesSurge capacityDriftLatency under loadCost per decisionHandover
01
Eval gateGate deployment on option-quality and entitlement-accuracy regression.
02
Surge testReplay a severe disruption day and hold the latency budget.
03
Drift watchDetect drift after a schedule or fleet change.
04
Cost per decisionModel and reduce cost per recovery decision at surge volume.
05
HandoverProduce runbook, legality regression suite and entitlement audit trail.
Assessment 08 — Production & surge
Two capstones — 20 hours
End-to-end agent design, development, deployment and testing, 10 hours each. Reviewed by a Senior SCDAI Engineer against a published rubric.
Capstone A — Disruption recovery agent
Build an agent that generates ranked recovery options for a replayed severe disruption, with cost and passenger impact quantified and crew-illegal options architecturally unreachable. Hold a control-centre latency budget under surge and measure option quality against a controller-labelled set.
Capstone B — Fare rule and entitlement agent
Build retrieval over filed fare rules and passenger-rights regulation that answers what is permitted and what is owed, citing rule, article and version, with reasoning an auditor could follow. Measure accuracy against an expert-labelled question set.
What you deliver, and how it is marked
- A running agent, deployed on the program's target stack, reachable by a reviewer.
- An architecture note: boundaries, tools, autonomy level and the trade-offs you took.
- A versioned eval set with a baseline and your final scores.
- A trace walkthrough of one success and one deliberate failure.
- Cost per task, measured — not estimated.
- A runbook and rollback path a client team could operate on Monday.
Problem fit15%Is this a decision worth automating, scoped to a defensible boundary?
Architecture20%Right pattern for the constraints; autonomy earned, not assumed.
Grounding & accuracy20%Measured groundedness against the eval set, with the failures named.
Safety & compliance20%Domain controls actually enforced, and demonstrably so.
Operability15%Traces, runbook, rollback, cost — could someone else run this?
Communication10%Can a client executive follow the reasoning in ten minutes?
One lab, start to finish
Every lab in this program follows the shape below. This is lab 05 in full — the brief you are given, the environment that is provisioned for you, the code you start from and the assertions that decide whether you passed.
Module 05 · Disruption agent design
Recovery options under a hard constraint
Generate rebooking options a controller can act on, with crew-illegal ones unreachable.
The brief
A hub weather event has cancelled six departures. You are given the affected passengers and their onward connections, the remaining schedule with seat availability, crew duty state, and the flight-time limitation rules. Return ranked recovery options with cost and passenger impact. The test is not whether you find a good option — it is whether an illegal one can be produced at all.
Provisioned for you
- 6 cancelled departures, 840 affected passengers
- Remaining schedule with seat availability
- Crew duty state and FTL rules
- Labelled key: 3 controller-accepted option sets
You start from this
def recovery_options(cancelled, passengers, schedule, crew, ftl):
"""Return ranked options: [(reassignments, cost, pax_impact, crew_plan)].
Legality is not a ranking input. Build the feasible set first, then
rank inside it: an illegal option must not exist to be ranked.
"""
# 1. build the legal crew-and-aircraft feasible set FIRST
# 2. only then assign passengers and cost the options
# 3. keep parties on one booking together unless splitting is unavoidable
raise NotImplementedErrorWhat the grader asserts
- No returned option breaches flight-time limitations, on any input including the adversarial set
- Legality is enforced by construction, not by filtering or ranking after generation
- Parties on a single booking are not split unless the option states why it was unavoidable
- Cost and passenger impact are quantified per option, not merely ranked
- Overlaps with at least 2 of the 3 controller-accepted option sets
05
Assessment
Assessed, not attended
Every module ends with a timed, randomised assessment delivered through the SCIKIQ Agentic AI Playground. The certificate requires a pass on all of them plus two reviewed capstones.
Assessment blueprint
What each test covers, how long it runs, and how many items are currently in the versioned bank behind it.
| # | Assessment & coverage | Items | Time | In bank |
|---|---|---|---|---|
| 01 | Operating model & scopingDelivery mandateAirline economicsValue sizingThin slicing | 25 | 35 min | 3 |
| 02 | Airline data landscapePSS dataDelay codingCrew dataFare formats | 25 | 35 min | 2 |
| 03 | Context engineeringCode resolutionTime handlingStructured outputsRefusal | 25 | 35 min | 2 |
| 04 | Retrieval & groundingFare rulesRegulation retrievalInterlineRetrieval metrics | 25 | 40 min | 2 |
| 05 | Agent design for disruptionConstraint handlingSolver integrationOption rankingLatency | 25 | 40 min | 2 |
| 06 | Systems integrationMCP designReal-time feedsApproval gatesDegraded mode | 25 | 35 min | 2 |
| 07 | Safety, rights & securitySafety boundaryLegality assuranceRights complianceLLM security | 30 | 45 min | 2 |
| 08 | Production & surgeEval gatingSurge capacityDriftCost per decision | 30 | 45 min | 2 |
Single best answerOne defensible option among plausible distractors drawn from real field mistakes.
Scenario judgementA client situation with constraints; you pick the action a Senior SCDAI Engineer would take.
Architecture selectionGiven non-functional requirements, choose the pattern and justify the trade-off.
Failure diagnosisA trace, an eval report or a cost curve — identify the root cause.
- Each module ends with a timed assessment; the program certificate requires a pass on every one.
- Items are drawn at random from a versioned bank, so no two learners sit an identical paper.
- Pass mark is 70%. Two retakes are included, each with a fresh draw.
- Scenario items carry double weight — they are the ones that predict field performance.
- Capstones are reviewed against a published rubric by a Senior SCDAI Engineer, not auto-graded.
Try four questions from the bank
Real items, drawn from 17 in this program's bank — weighted toward the scenario and diagnosis types, because those are the ones that predict field performance. Instant feedback, nothing saved.
Four sample items — one attempt each, then the reasoning is shown.
Q1A carrier asks for "AI for customer experience". Where does the value concentrate?
Disruption is where an airline spends money it did not plan to spend and loses customers it already won. It is also where decisions are fast, constrained and currently manual.
Q2A user asks an agent whether a crew pairing is legal. What should the agent do?
Legality is a regulated determination with a system of record. An agent that estimates it will eventually be wrong in the direction that grounds an aircraft or breaks a rule.
Q3How should crew-illegal recovery options be handled by an option generator?
A hard constraint enforced by ranking is a hard constraint that will eventually be violated under time pressure. It has to be structurally impossible.
Q4A source system fails mid-recovery. What should the agent do?
Silent degradation is the dangerous option: the controller keeps trusting output that is now based on stale inputs. Declaring the degradation preserves the trust.
06
Career & outcomes
What this makes you, and where it goes next
The same ladder whichever specialisation you enter through — what changes is the domain you go deep in. Below: how the program is delivered, the skills it moves, the roles it leads to, and the specialisations closest to this one.
Available two ways
The same labs, assessments and capstones, delivered to an enterprise cohort or to individual professionals.
B2B
Direct to enterprise
Cohorts of 20 to 2,000+ on your own tenancy, with your data patterns and your cloud. Skill-gap baselining up front, per-team mastery reporting throughout, and capstones scoped against your real backlog so the output is deployable work.
- Private cohort on your tenancy
- Baseline skills assessment before kick-off
- Capstones scoped to your backlog
- Per-team mastery dashboards for L&D
- Optional Senior SCDAI Engineer architecture reviews
B2C
Direct to individual professionals
The same labs, assessments and capstones for individual engineers and analysts, run on shared infrastructure with a fixed cohort calendar. You leave with a graded portfolio, not a certificate of attendance.
- Fixed-date cohorts, self-paced within them
- Same graded labs and item bank
- Portfolio-ready capstones you own
- Certificate on a pass in every module
- Community and office-hours access
Skill matrix
Find your row and aim one column right. The Playground scores you against this after every module.
| Skill | Beginner | Intermediate | Advanced |
|---|---|---|---|
| Airline domain fluency | Knows the segments. | Maps schedule decisions to cost and revenue. | Sizes disruption and ancillary value credibly. |
| Operational data | Reads a movement feed. | Assembles a connection graph that matches reality. | Designs for surge, degraded mode and inconsistent delay coding. |
| Constraint engineering | Aware of crew legality. | Makes illegal options unreachable at the tool layer. | Runs a legality regression suite as a release gate. |
| Context engineering | Writes clear prompts. | Structures retrieval, tools and state deliberately. | Designs context strategy for reliability and cost at scale. |
| Retrieval & grounding | Builds basic vector search. | Tunes chunking, hybrid search and reranking. | Designs graph + vector grounding with measured recall. |
| Agent orchestration | Runs a single tool-calling agent. | Builds supervised multi-step and multi-agent flows. | Designs autonomy boundaries and failure containment. |
| Tool & system integration | Calls a documented API. | Writes an MCP server over a system of record. | Designs a least-privilege tool estate across systems. |
| Evaluation | Eyeballs outputs. | Builds labelled eval sets and regression gates. | Runs online evals with drift and judge calibration. |
| Observability & cost | Reads logs. | Traces runs, tracks tokens and latency. | Owns cost per task and capacity planning in production. |
| Security & guardrails | Adds output filters. | Mitigates the OWASP LLM Top 10 in a build. | Threat-models an agent estate and proves controls. |
| Client delivery | Takes notes in a workshop. | Runs discovery and scopes a thin slice. | Owns the account technically, from scope to handover. |
Where it takes you
The SCDAI ladder is the same whichever specialisation you enter through — what changes is the domain you go deep in.
01
Data and AI EngineerEntry
Python, LLM APIs, first RAG build, prompt and context basics.
02
Associate Data and AI EngineerPractitioner
Client-facing discovery, tool integration, agent build under supervision.
03
SCIKIQ Certified Data and AI EngineerSpecialist
Owns a deployment end to end — scoping, build, evals, handover.
04
Senior Data and AI Engineer / Domain LeadSenior
Multi-agent architecture, regulated deployments, reference patterns for the practice.
05
AI Architect / Practice LeadLeadership
Portfolio of accounts, delivery standards, hiring and capability strategy.
Terms worth knowing
- IROPS
- Irregular operations — disruption, and where airline cost concentrates.
- PSS
- Passenger Service System — reservations, ticketing and departure control.
- Fare rule
- The filed conditions governing what a fare permits, in a dense positional text format.
- Flight-time limitation
- The regulated cap on crew duty and flying hours — a hard constraint.
- Misconnect
- A passenger who cannot make their connection, and the unit disruption is measured in.
- Ancillary
- Revenue beyond the seat: bags, seating, lounge, change fees.
- Interline
- Carriage involving another carrier under an agreement between them.
- RASK
- Revenue per Available Seat Kilometre — the headline commercial metric.
- SCDAI
- SCIKIQ Certified Data and AI Engineer — the credential this program awards.
- SCDAI engineer
- An engineer embedded with the customer who turns their constraints into a working system.
- Thin slice
- The narrowest end-to-end path through a workflow that still produces business value.
- Agentic RAG
- Retrieval where the agent decides when, what and how often to retrieve.
- MCP
- Model Context Protocol — open standard connecting an agent to tools and data.
- A2A
- Agent2Agent — protocol for agent-to-agent delegation across vendors.
- Eval set
- A labelled, versioned set of cases an agent must pass before release.
- LLM-as-judge
- Using a model to score outputs, calibrated periodically against human ratings.
- Guardrail
- An enforced constraint on an input, output or action — not a prompt request.
- Human-in-the-loop
- A required human approval step before a consequential action executes.
- Cost per task
- Total token, tool and compute cost of one completed unit of work — the engineer's unit economic.
- Golden path
- The reference architecture a practice standardises on so deployments stay reviewable.
- Handover
- The point at which the client can operate, evaluate and change the system without the engineer.
Related programs
Skill a team, or join a cohort
B2B cohorts run on your tenancy with capstones scoped to your backlog. B2C cohorts run on a fixed calendar.
07
Enrolling
What you need, what you get, and what it costs
Stated plainly enough to rule yourself in or out without a sales call: the prerequisites, how the program runs, exactly what the credential is worth, and the questions everyone asks.
Before you start
Stated plainly so you can rule yourself in or out without a sales call. Nothing here is a formal qualification — it is what the first lab assumes you can already do.
Technical
You should already be able to do these
- Comfortable in Python: functions, typing, virtual environments, reading a stack trace
- Have called an LLM API and handled a structured response
- Git, the command line, and reading someone else’s code without panic
Domain
What we assume, and what we teach
- Know the difference between a booking and a flight; the rest of the network is taught
- Willingness to treat crew legality as a hard constraint rather than an optimisation
Time and access
What the program asks of your week
- A machine that can run a browser and a terminal — labs run in the Playground, not locally
- Roughly 8 hours a week for the taught weeks, plus 20 hours for the two capstones
How it runs
Cohort dates and pricing are confirmed on enquiry rather than printed here, because both move with the intake.
- Format
- Self-paced within a fixed cohort window, delivered in the SCIKIQ Agentic AI Playground
- Taught hours
- 80 hours across 8 modules
- Capstones
- 20 hours — two end-to-end builds, reviewed not auto-graded
- Typical duration
- 10–12 weeks at around 8 hours a week
- Level
- Intermediate → Advanced
- Built for
- Engineers and analysts deploying AI across airline commercial and operations
- Next cohorts
- — confirmed on enquiry
- Pricing
- — confirmed on enquiry
The credential
The credential is awarded per specialisation, so it names the domain or stack you were assessed in rather than claiming general competence. On this program the badge reads SCDAI — Airlines.
How it is earned
- A pass at 70% or above on every module assessment — not an average across them
- Both capstones reviewed against the published rubric by a Senior SCDAI Engineer
- All labs submitted and machine-graded green
- Awarded as
- A verifiable digital badge with a public credential record
- Valid for
- Two years from award
- Recertification
- One current-year assessment plus a refreshed capstone
- Retakes
- Two included per assessment, each with a fresh item draw
- Named specialisation
- The badge states the domain or stack, e.g. SCDAI — Healthcare
- Verification
- Third parties confirm a holder from the credential record, no login required
A certificate that cannot be checked is decoration. Every award resolves to a record showing the specialisation, the award date and the assessments passed.
The certificate you leave with
This is the credential the program awards, shown exactly as it is issued — with the specialisation named, the assessment record attached and a verification link anyone can check without an account.
SCIKIQ Data Academy
Certificate of Professional Certification
This is to certify that
Your name
has been assessed and certified as
SCIKIQ Certified Data and AI Engineer
Airlines
- Specialisation
- Airlines
- Assessments passed
- 8 of 8, each at 70% or above
- Labs graded
- 40 of 40 machine-graded green
- Capstones reviewed
- 2, against the published rubric
- Taught hours
- 80 across 8 modules, plus 20 capstone hours
AwardedOn completion
Valid untilTwo years from award
Credential IDSCDAI-AIRLINES-<issued>
Senior SCDAI Engineer, Certification Board
- 01The specialisation is on the certificate
Not “Data and AI Engineer” but the domain or stack you were actually assessed in. A general claim would be a weaker one.
- 02The assessment record travels with it
Modules passed, labs graded and both capstones reviewed — so the credential states what was measured rather than that you attended.
- 03Anyone can verify it, without a login
The credential ID resolves to a public record showing the specialisation, the award date and the assessments passed.
Share it on LinkedIn
Add it to your LinkedIn profile in one step. The link pre-fills the certification fields from the credential record, so the entry on your profile matches the record a reader can check.
Add to LinkedIn profile The button is live on your real certificate; here it opens LinkedIn pre-filled with this specialisation so you can see exactly what the profile entry will say.A certificate that cannot be checked is decoration. Every award resolves to a record showing the specialisation, the award date and the assessments passed.
Questions we get asked
The objections that come up in every conversation about this program, answered without the brochure voice.
Is this a course or a certification?
Both, and the second is the point. Eight timed assessments and two reviewed capstones stand between you and the credential, so a pass means someone measured the skill rather than recorded your attendance.
How is it self-paced without falling apart?
Every lab is provisioned, graded and unblocked by the Playground rather than by an instructor. That is what lets a cohort of 2,000 cost the same faculty time as a cohort of 20 — and why you are never waiting on someone to mark your work.
What if I fail an assessment?
Two retakes are included per assessment, each drawing a fresh item set from the bank, so a retake is a genuinely new paper rather than the same questions again.
Do I need my own cloud account?
For the domain specialisations, no — labs run in provisioned sandboxes. For the four tech-stack programs you will want access to that platform, since deploying into a real subscription is much of the point.
Can we run this privately for our team?
Yes. B2B cohorts run on your own tenancy with your data patterns, a skills baseline before kick-off, per-team mastery reporting, and capstones scoped against your actual backlog so the output is deployable work rather than an exercise.
Which specialisation should I take?
Take the domain you deploy into. If you move across industries, take a tech-stack program instead and pick up domain context on the engagement. The chooser on the programs page will narrow it.
How current is the material?
The trends, platform capabilities and regulatory positions are reviewed each quarter, and every external claim on these pages links to its source so you can check the date yourself.
What do I actually leave with?
A graded portfolio: forty machine-graded labs, two reviewed end-to-end agent builds with measured evaluation and cost per task, and a verifiable credential naming your specialisation.
Still deciding?
Tell us the systems you deploy into and we will say plainly whether this specialisation is the right one — or which of the 21 is.